The proposed data bases for stroke and traumatic coma will provide data that require both critical application of existing statistical methodology and development of new statistical techniques in order to extract meaningful information and to draw valid conclusions from the accumulated data. A continuing review of the statistical activities by existing medical data bases and a search of the relevant literature was initiated. Some of the major problems and associated methodologies that have been formulated are: validation and quality control of the data, analysis of variation due to observers, centers, and patient cohort characteristics; association of indicators and treatment with patient outcome, and analysis of trends in time oriented data. A set of descriptive and associative statistics are available for the analysis of symptoms, indicators, and outcomes. Problem areas that have been identified and that require additional study are: recognition of and adjustment for biases, derivations of life stables with competing risks and utilizing concomitant information, and the use of stratification and how it affects analysis.